I'm pretty stoked about this. I've been following Rust since 0.4, and it's been super fun to play around with. I have two major problems with it, but none of them are Rust's fault -- purely flaws of my own:
1) I have no idea what to do with it. I'm primarily a web developer, and I really want to get into systems programming, but I have no idea what to do with it or where to start. I wound up implementing about 50 different Rosetta Code things in Rust 0.9 and 0.10, but I still never came up with a really solid use-case that I could start and finish.
2) Given that the last time I touched systems programming was in college (15 years ago), I've found it difficult to wrap my head around some of the concepts (ownership in particular). I actually hung out in the IRC chat room for awhile, but I felt more than a few times that people were treating me like an idiot because of my questions ("Ha! You did X? Why in the world would you do that??"), so I wound up not really going back. I'm sure it was all my imagination and that I'm just being sensitive, but I think it would be very helpful to have more people interested in getting new people on board -- specifically people like me, with over a decade of programming experience in higher-level languages.
I guess the TLDR of this is: I love Rust, but I don't know what to do with it.
Congrats to the Rust team, though. 1.0 is a huge milestone. Good job!
I'm in your boat, I'm a web developer who's always been curious about systems programming and has been following Rust as a way to get into the domain without all the tedious runaround and rampant paranoia associated with manual memory management. But, I still don't know what I should actually do with this knowledge now. :)
> I actually hung out in the IRC chat room for awhile,
> but I felt more than a few times that people were
> treating me like an idiot because of my questions ("Ha!
> You did X? Why in the world would you do that??")
Unacceptable. I'm an op in #rust (bstrie), so feel free to PM me their nicks and I'll see to it that they treat new users better in the future.
As a counter-anecdote, I'm in a very similar boat as the parent comment, and I have been astounded how consistently all the people in irc who definitely could make me feel like an idiot are unwilling to do so. It is a really refreshing experience after years of going into rooms full of arrogant people who actually don't know very much, as opposed to Rust's irc where everyone seems to know everything and are humble nonetheless.
I don't disagree! The vast majority of the IRC group was very helpful! I should have mentioned that in my original post. There was just one particular day where I was struggling with a concept and I felt like people were making fun of me and it just really made me not want to touch Rust for a few months.
But that's not really the point. I think the point is that there's a need for a sort of onboarding process for people who have never had to deal with some of these basic concepts. For instance, in web development, there's not a lot of need to know much about ownership or memory. The popular languages (Java, PHP, Python) kinda handle those things for you. So when you dig into Rust, you spend a lot of time thinking things like "wtf is a pointer?" and "wtf is a reference?" instead of actually building something useful.
That's a really frustrating part for me. I can build anything I want in PHP or Python or Java. I know those languages like the back of my hand, and if you give me a problem, I can solve it.
But then I start with a language like Rust, which concerns itself with all kinds of things I've only vaguely heard of, and then I'm completely lost. Super, super, super frustrating. It's a struggle to do a basic blackjack game -- something I could do in my sleep with Python!
So you don't know these concepts, which is already frustrating, and then you go into chat and you say "WTF is a pointer?" and hopefully people say "Oh, dude, it's so easy, imagine. you have this..." instead of "That's basic computer science, man. RTFM."
To be fair, I think Rust's documentation back at 0.9 was great. I'm positive it's way better now. But back at 0.9, it didn't answer the questions I had, which is why I had to go to chat in the first place.
But yes, most people were very, very helpful and I really did learn a lot. :)
If you're still stuck on pointers (although sounds like you eventually got past it), Kernighan and Ritchie have a very no-nonsense, straightforward presentation of the concept.
Of course, the basic concept of a pointer is not really that tricky. It's how to use them, what you can do with them, and how to avoid accidentally shooting off your foot with them that are the tricky parts.
Anyhow, if you know any other programmers who want to understand pointers, I don't think you can go too far wrong pointing them towards Kernighan and Ritchie.
I, for one, found the arrays and pointers chapter in K&R Edition 1 to be mostly useless (per my other comment in this subthread I used the Lions Notes). Do you know if it was significantly improved in the 2nd Edition you recommend?
I happen to have both editions handy. Some sub-sections seem to be 30%-50% rewritten. A big change is there are lots of graphics. 2e seems to be improved, but it's hard for me to estimate how much of an improvement it really is.
I scanned the first couple of pages again in the PDF, and still like the simplicity of the presentation, avoiding indirect metaphors ("a pointer is kind of like...").
I never got a copy of Lions, unfortunately. Seems a shame to not have more books in that style.
Hmmm, I read the first N pages of 2nd Edition of the chapter and it sure looks clear and good to me, and one area where I found it confusing looks like it was improved (but who can say 33 years later?). I won't make an anti-recommendation of this edition, but would be curious if anyone has problems with it.
So one obvious solution is to annotate a body of real code ... but in general you're suffering from trying to learn a moving target, and good documentation tends to trail....
That Version 6 UNIX(TM) was updated/reimplemented as Xv6 https://en.wikipedia.org/wiki/Xv6 ... maybe someone should do a pedagogical Rust version??? Maybe x86_64 and/or ARM?
"it baffles me people don't get pointers nowadays"
If you continue to be baffled in the face of consistent evidence then the problem is probably in you rather than in reality (or you're not actually baffled, but instead are feigning confusion as a way of projecting superiority). Honestly what's there to be baffled by? There are many programming environments that are both common and quite productive that have no pointers at all. I've gone literally years of my working life without touching a single pointer in production code.
That's an unusually harsh response I know, but I find this sort of attitude incredibly toxic, and it has hollowed out several communities I've been adjacent to. Some people know some things but not others. You can make them feel crappy by signalling that they are damaged or deficient, or your attitude can be positive and inviting.
> I've gone literally years of my working life without touching a single pointer in production code.
But you did use references, right?
What baffles me is how many young developers code without caring about the underlying infrastructure. Not caring to understand what makes their coding possible.
I have met many that deal with programming like car owners that just want to drive down to the supermarket, so to speak.
> You can make them feel crappy by signalling that they are damaged or deficient,
I explicitly stated I wasn't targeting the OP with my comment.
> or your attitude can be positive and inviting.
That is why I pointed out a few, quite good, books where the subject of reference vs pointers happen to be described quite well in a programming language that isn't as intimidating as C or C++, while allowing for the same type of coding.
One of them even shows how to build an entire computer.
A couple other people have jumped on this, but it's also worth pointing out the specialization that goes on in the field now.
For example, I'm almost exclusively writing web applications, and have been for years (coming up on 15 years soon).
Granted, I do know a bit of C and some other more systems-y languages (and more than a little bit about the underlying implementation of Python, which is the language I prefer to use), and I even know a fair amount of underlying fundamental theory, but... ultimately my brain is of finite size and I have only finite time. So I have to prioritize like everyone else, and choose what I will spend time learning and what I will devote brain cells to remembering.
The result of this is that there are things which, to someone outside of web development, would seem horribly arcane and irrelevant, but which I have to know, and as a result there are things those other people have to know which, to me, are horribly arcane and irrelevant.
And the further result is that when I see people wading into web development from other fields of programming, I often see them make mistakes -- sometimes big, security-hole-opening, money-losing mistakes -- that are on par with someone blundering into C with no knowledge of pointers, memory management, bounds-checking, etc.
So just remember that what seems hugely important and required-learning to you may not be to someone else, and that something which seems irrelevant to you may be hugely important required learning to them :)
The link to reference material was nice but what I'm hoping to draw attention to is about about tone. Your tone is scornful, judgmental, and aloof. If we worked together and I had a question on a subject of your expertise I might think twice before consulting you for fear of being dressed down publicly for my ignorance.
"What baffles me is how many young developers code without caring about the underlying infrastructure. Not caring to understand what makes their coding possible."
Our industry has many layers of abstraction, and there's always something to be gained by exploring more of the space (e.g. application domain, high level language, low level language, compiler, compiler IR, assembly, process, os, hardware apis, hardware implementation, electrical engineering, theoretical physics). However it is destructive to the social dynamics of a community when members scoff at or impugn the motives of those who do not know what they do. Both sides probably know things that the other doesn't too, why not learn from one another?
If instead of being content with being baffled with these colleagues of yours who are incurious about pointers, why not show them something cool that you can build with them! E.g. walk them through that elegant implementation of a hashmap in C that uses pointers to pointers from the front page of HN a couple of days ago.
Just saying you're not targeting someone doesn't stop them from feeling bad about not understanding something that is apparently obvious.
You have a different background to many programmers, is it unreasonable that they don't understand/know something that you do? It certainly doesn't make them stupid and statements like "baffled that people don't understand pointers" definitely imply that they are somehow deficient.
In any case, why does everyone have to care about in inner workings of everything they use? It seems perfectly reasonable for computers and programming languages to be things that people can use without having to understand the details, like cars, microwave ovens, aeroplanes, TVs, radios, analog watches, GPS (the list goes on).
> That is why I pointed out a few, quite good, books where the subject of reference vs pointers happen to be described quite well in a programming language that isn't as intimidating as C or C++, while allowing for the same type of coding.
Pointing out books is nice, but it doesn't make up for effectively abusing people who might need them: "read this to learn more" is way more pleasant than "I can't believe people don't know this, read this to learn more".
What baffles me is the number of programmers who use technology like cars, cell phones, fridges, planes, trains and advanced medicine without understanding what makes the underlying technology possible.
You don't do processor microcode or memory controller code, why should you care about that? People who work in managed languages don't care about pointers. I don't remember which register stores the result of a division either. Why should I care about it?
Except Java hasn't had them for 20 years and most Java programmers don't even think about how it takes 4 levels of indirection to just read something that would be a struct in C.
> There are many programming environments that are both common and quite productive that have no pointers at all.
This is a bit misleading. The vast majority of languages have pointers (languages without object identities exist - functional languages for example - but they are relatively the exception). However, those languages tend to give them other names.
What non-systems languages lack is pointer arithmetic and native notions of pointer-to-pointer (though it is always possible to simulate the latter).
I agree the person you responded to showed a toxic attitude. I believe/hope that the first step to improvement for people who think they don't get pointers is for them to realize they've been working with pointers all along.
The concept of references was introduced in my introduction to Java course I guess at the same time that the concept of objects were explained more fully. At least in Java, I feel that you have to understand the concept of references, to be mindful of aliasing, the fact that primitive values can't be aliased (I've tried twice to make a method to swap two primitive values, of course failing at that), and so on. Pointers don't really feel like a conceptual jump, coming from Java - the fundamental concept of indirection is still there. The fact that you can do pointer arithmetic and such on them, while error prone and harder to deal with than references, did not feel that hard to understand, conceptually. Granted, I have very little experience with C-level programming, so I be missing some subtle points.
There might be a difference for people who are used to languages which are more consistent than Java (like: "all values are really objects"). Do people not have to know about references in languages like for example Python?
This may largely be because we don't actually know everything, thanks to the fact that Rust changes so quickly that even the experts must constantly re-learn the language. :) But I'm very invested in maintaining that atmosphere of respect as we progress towards stability.
For 1, it's true that it's not clear Rust is going to be great for web development. http://arewewebyet.com sums things up. That said, Rust does run the backend of Crates.io, and I've been toying around with it as an Ember backend, so we'll see :)
2) I'd love to hear your thoughts on my new ownership guide, which landed last weekish: http://doc.rust-lang.org/guide-ownership.html thoughts very welcome, either here, email, or on the issue tracker.
Ownership is the one concept in Rust that hasn't really existed at the language level before, so it's the most unusual thing to grok.
> I felt more than a few times that people were treating me like an idiot because of my questions
Sorry to hear that. :/ We try to keep it 100% welcoming, but sometimes, people slip through the cracks.
> I've been toying around with it as an Ember backend
Care to divulge some of your experiences so far? Like others here, I'm a webdev searching for a point of entry in Rust, and seeing that there's web frameworks out already I thought they'd might work. I want to start a side project with a Backbone/Marionette front but have been wondering whether to use Rails (day job), Node, or now even Rust. Thanks.
Just saw your rustforrubyists...definitely going to follow that!
I haven't had the time to make it great, just hacking away.
> Just saw your rustforrubyists...
I appreciate the sentiment, but you probably shouldn't bother. Once I started working on the Guide, I stopped doing more than minimal upkeep on R4R. I consider it deprecated at this point. :)
I don't necessarily need web-based ideas, but I don't know enough about the systems programming ecosystem to have an idea of where to start! A text editor? A game? A virtual planetarium? Machine learning? I have no idea! (Like I said, not your/Rust's problem...)
There are a lot of web-based tools that have C implementations, so fun projects may include writing pure-Rust implementations of these. Imagine, say, the CommonMark markdown parser (https://github.com/jgm/CommonMark) implemented in Rust. Making it as fast as the C version would be a challenging optimization problem, but I bet it's doable. :)
I feel like the projects where where rust is best positioned in terms of comparative advantage fall into two categories:
1) Large and complex systems that are slow to evolve due to being written in languages that are unsafe or are relatively poor in abstraction (in addition to the immense domain complexity). Examples: SQL and/or distributed databases, Paxos-ish systems, Operating Systems, IDEs, GIS, etc.
2) Extremely high performance systems on the fringe of research. Examples: MIP/LP/CP/SAT Solvers, Theorem Provers, Formal Verification tools, Rete-based Rule Engines, Machine Learning algorithms and frameworks, Statistical Samplers for MCMC, etc.
Unfortunately (maybe, I don't really know if it will turn out well or not), most of the hype for rust comes from back-end web developers. That is fine...I'm sure there is some performance to eek out there...but it is also trying to compete in an area where garbage collection has its greatest competitive advantages (namely, long running server processes). The comparative benefit is relatively small.
> most of the hype for rust comes from back-end web
> developers
It may seem that way because 1) most of the hype for anything programming-related these days comes from back-end web developers, and 2) because back-end web developers are the primary audience of HN. At Rust meetups that I've attended I've seen interest in people doing video games, robotics, AI, numerical work, databases... they just don't have as public and excitable a forum as this. :)
I honestly wish we could get some of HN's mobile crowd hyped on Rust. A Rust-y take on Xamarin's ideas would be phenomenal, and with resource constrained phones, the performance gain from not having garbage collection or reference counting would be certainly welcome.
I suspect that crypto libraries and embedded systems might also be good areas for Rust. Basically anywhere that a buffer overflow or memory leak is a relatively high-penalty failure.
> compete in an area where garbage collection has its greatest competitive advantages (namely, long running server processes)
I thought it was the opposite that was true. A garbage collector can/will leak memory over time (days) and will use up to 6-8 times more memory and can have performance problems etc. All long running server processes I know of (database, mail, http servers) manage their own memory. Exceptions are VM based systems and it is not unheard of to have a trivial Tomcat app use up to 1GB of memory.
Here, I come from the Lisp world where PDP-10s had magnetic core memory, and a max of 1MiB (of 9 bit bytes) in address space. Lisp Machines used DRAM, but at least early on less than an address space (64MiB of 8 bit bytes), and keeping your working set inside that was an issue if for no other reason than GC by the time I showed up was incremental copying without generations (i.e. an increment of GC would tend to trigger a page fault).
So I assure people like sovande that this can be done correctly, even if it's convenient to "ship it now!" and later find out what's "leaking" (keeping unneeded stuff from the tender mercies of the garbage collector).
Think client side native apps. Like a cross platform layer for a mobile/desktop app suite. With web dev, you have the advantage of being able to accept the JVM's relatively long spin up time and it's resulting speed over time. You have the flexibility to use things such as clojure or erlang. Client apps can't accept that kind of restriction very well.
I know of many client side libraries that are in C++ because it works everywhere(tm). Being able to maintain them in something like rust would be far nicer.
Right now they are failing to build due to a recent breaking change, which made structs not implement Copy by default any more. The library is in active development, so that should be fixed soon enough.
Web developer here, i just srarted to write a interpreter for a small lisp, granted not system programming but I can still use the low level features to my advantage.
Its pretty fun, but im looking forware to stability too.
Writing a Lisp "in a week" (that would be putting most of your time into the project) is a classic thing to do, and while it indeed isn't exactly system programming, if you keep it simple (e.g. maybe limit it to atoms, lists and fixnums), but include precise GC, you'll learn a lot of things you need.
Hmmm, I've been wanting to learn Rust, but I always need a project to learn a language, and I haven't written a Lisp "in a week" yet ^_^. Yeah, thanks for the idea.
Well the project actually goes beyond just that, its a bytecode interpreter and I want to experiment with a JIT (something I have wanted to do for years).
Sadly its not working currently, I started a new job and Im moving so I had very little time the last 2 weeks and I didn't manage to keep up with the changes in rust.
You're a web developer, and you don't know what to do (in terms of how to spend your time) with a language geared toward systems.
That makes a lot of sense! At a certain scale, they're different types of work. It's like saying, "I'm a race car driver with a fighter jet. I've always admired them, but I'm just not sure where to fly!"
These types of languages, and more broadly any compiled language like C or C++, are very well suited to solve problems that your dynamic or scripted language can't solve at scale. Say, for example, as a web developer you built a website on Ruby on Rails that converts user-uploaded photos to morph, based on some algorithm, into a cute, unique photo of a kitten.
You have several concerns here:
- Processing the image, hopefully in the background so it's non-blocking and you can control resource needs.
- Storage. Is it on a local disk? S3? Something close by but on the network? If you have a performance bottleneck, is I/O an issue?
- Memory/CPU. How many resources are needed for n- jobs? Is it starving resources from serving additional web requests (really a separate concern, when you think about it!)
I would say for these types of problems (and this is admittedly a contrived example), Rails + some Gem or even pure Ruby might just absolutely suck and eat it.
These specialized languages can be a super useful tool for figuring out "simple problems with complicated solutions." They also offer the comfort of compiling down to a binary that might have no external dependencies (so, very portable depending on the compiler/runtime, and probably very reliable).
Is that the right thing for a pure web developer to play? Depends on the developer, but it's not surprising you're not quite sure what to do with it. It's possible you just don't have the right type of problem to solve yet.
That being said, Rust or Go are still fun to play with! Good for you for exploring.
That's exactly what I mean. :) A lot of the things I rely on for web development are so advanced under the hood that I don't even know where to begin. Like, let's say I want to write a tool that will let me take photos, resize them, crop them, add watermarks. I don't even... where do you start?
I wrote an LZW compressor in Rust 0.9 way back when, and it sucked pretty badly, but I wouldn't even know how to open a JPG or PNG, much less the mystical art of resizing it!
Call up graphicsmagick - either through spawning processes or binding to the library. All of those are solved problems, so if you don't have time/inclination to learn how they're solved, just use something that already did :)
> 1) I have no idea what to do with it. I'm primarily a web developer, and I really want to get into systems programming, but I have no idea what to do with it or where to start.
I'm thinking profiling some stuff and writing native accelerators, in Rust instead of C. More of an excuse than a need, but hey.
> I've found it difficult to wrap my had around some of
> the concepts (ownership in particular).
Have you tried reading the Ownership Guide (http://doc.rust-lang.org/guide-ownership.html)? I actually haven't read this particular guide myself (it didn't exist when I got started with Rust), but it sounds relevant :)
My main use for C code is embedded / bare-metal programming, so I was curious if Rust could run directly on ARM boards. I found https://github.com/hackndev/zinc but haven't given it a try yet.
I have found the IRC channel to be an incredibly supportive and helpful place, despite my barrage of dumb questions. I've been on IRC since the mid 90s and never seen anywhere so friendly, welcoming, and tolerant.
1) I have no idea what to do with it. I'm primarily a web developer, and I really want to get into systems programming, but I have no idea what to do with it or where to start. I wound up implementing about 50 different Rosetta Code things in Rust 0.9 and 0.10, but I still never came up with a really solid use-case that I could start and finish.
That said, I've been really impressed by rust-doom (https://github.com/cristicbz/rust-doom) and the new Iota editor (https://github.com/gchp/iota), and I wish I had the know-how to even start down that path.
2) Given that the last time I touched systems programming was in college (15 years ago), I've found it difficult to wrap my head around some of the concepts (ownership in particular). I actually hung out in the IRC chat room for awhile, but I felt more than a few times that people were treating me like an idiot because of my questions ("Ha! You did X? Why in the world would you do that??"), so I wound up not really going back. I'm sure it was all my imagination and that I'm just being sensitive, but I think it would be very helpful to have more people interested in getting new people on board -- specifically people like me, with over a decade of programming experience in higher-level languages.
I guess the TLDR of this is: I love Rust, but I don't know what to do with it.
Congrats to the Rust team, though. 1.0 is a huge milestone. Good job!